Loading...
Loading...
Google announced the Gemini 3.5 family with Gemini 3.5 Flash released publicly and Gemini 3.5 Pro used internally and expected next month. Flash targets agent workflows and complex, long-horizon programming tasks and is integrated across consumer apps, developer platforms, and enterprise tools. It offers massive context and output windows and a new Interactions API for server-side history, but comes with significantly higher pricing—roughly 3x–6x previous Flash rates—pushing costs near Gemini 3.1 Pro. The rollout highlights Google’s tiered model strategy, aiming to balance advanced capabilities with differentiated offerings for developers and enterprises amid rising LLM costs.
Google's Gemini 3.5 rollout affects product integrations, developer costs, and enterprise AI strategies as Google embeds higher-capability models across its ecosystem. Tech professionals must assess cost, performance trade-offs, and migration paths for apps and workflows.
Dossier last updated: 2026-05-21 09:16:51
Google announced Gemini 3.5 Flash at Google I/O and has deployed it broadly across consumer, developer, and enterprise products, including the Gemini app, AI Mode in Search, Google Antigravity, Gemini API in AI Studio and Android Studio, and Gemini Enterprise Agent Platform. The model (gemini-3.5-flash) supports a January 2025 knowledge cutoff, 1,048,576 input tokens and 65,536 output tokens, but lacks computer use. Google also introduced a beta Interactions API for server-side history management. Crucially, pricing jumped substantially—about 3x–6x versus earlier Flash variants ($1.50/million input, $9/million output), nearing Gemini 3.1 Pro rates—part of a wider industry trend of higher costs for top-tier models. Benchmarks show much higher run costs compared with prior Geminis and competitors.
@Balder13946731: Gemini 3.5 Flash, 仅仅是快速的版本就与深度思考的Opus 4.7 和 GPT5.5相当，那么毫无疑问 Gemini 3.5 Pro将强得可怕。 Antigravity和操作系统以及整个Google生态深度整合，无形之中与
Google on May 19 unveiled the Gemini 3.5 model family, debuting Gemini 3.5 Flash as the first public release and announcing that a higher-tier Gemini 3.5 Pro is in internal use and slated for public release next month. Gemini 3.5 Flash is positioned to excel in agent-based workflows and programming, focusing on tackling complex, long-horizon tasks that deliver practical application value. The staged rollout signals Google’s push to iterate on its large language model lineup with differentiated tiers for performance and capabilities, intensifying competition in advanced AI models and developer tools.
Google launched Gemini 3.5 Flash at Google I/O and made it generally available across consumer apps (Gemini app, Google Search AI Mode), developer platforms (Google Antigravity, Gemini API in AI Studio and Android Studio), and enterprise offerings (Gemini Enterprise Agent Platform). The model ID is gemini-3.5-flash with a Jan 2025 knowledge cutoff, 1,048,576 input-token context and 65,536 output-token max. Google also introduced a beta Interactions API for server-side history management. Pricing jumped substantially to $1.50/million input and $9/million output—3x–6x prior Flash variants—bringing costs close to Gemini 3.1 Pro and reflecting an industry trend of higher prices for newer LLMs. Benchmarks indicate much higher run costs versus earlier models.